Using Number Needed to Treat to Interpret Treatment Effect

Transcription

1 Continuing Medical Education 20 Using Number Needed to Treat to Interpret Treatment Effect Der-Shin Ke Abstract- Evidence-based medicine (EBM) has rapidly emerged as a new paradigm in medicine worldwide. The clinical medicine in twenty-first century could be the era of EBM. Randomized controlled trial has been regarded as the gold standard for evaluating the treatment effect of a new drug or a new therapy. The effect of a treatment versus controls may be expressed in relative or absolute measures. Relative measures include relative risk, relative risk reduction, and odds ratio. Absolute risk reduction and number needed to treat are absolute measures. For rational decision-making, absolute measures are more meaningful because they have taken baseline risk and the amount of clinical benefit into account. The number needed to treat (), the reciprocal of the absolute risk reduction, is a useful estimate of treatment effect. Interpreting a should be very cautious accompanied by information about the experimental treatment (including drugs and surgical procedures), the control treatment for comparison, the baseline risk of the study population, the length of the follow-up period, and an exact definition of the endpoint. Key Words: Evidence-based medicine, Randomized controlled trials, Treatment effect, Number needed to treat From the Section of Neurology, Department of Medicine, Chi Mei Medical Center, Tainan, Taiwan. Received January 7, Revised and Accepted January 25, Reprint requests and correspondence to: Der-Shin Ke, MD, PhD. Section of Neurology, Department of Medicine, Chi Mei Medical Center, No. 90, Chung Hwa Rd., Yung Kang, Tainan, Taiwan. dershink@ms7.hinet.net

2 2 (Number needed to treat) (2) () (The number needed to treat) Acta Neurol Taiwan 2006;5:2-26 trials, RCTs) (treatment effect) (randomized controlled (superior to) (no difference to) (gold standard) () (placebo effect) dershink@ms7.hinet.net

3 22 (clinical trial) (reporting) evidence-based medicine, EBM the Consolidated Standards of Reporting Trials (The CONSORT Statement) (3) 200 The CONSORT Statement (5,6) (4) RCTs (patient centered health care) (problem based learning) (clinical practice guidelines) RCTs The CONSORT Statement (number needed to treat, ) 60% (2) (treatment effect) (severe adverse effect) (primary endpoint) (secondary endpoints) (adverse effects) (impact factor) (event rates) (control event rate (experimental event rate EER) CER)

4 23 (measures) (estimates) () (confidence intervals) (relative effect) risk RR) (odds ratio OR) (7,8,9) () (relative (relative risk reduction ) (2) (absolute effect) (absolute risk reduction ) (risk difference) () (independent variable) (dependent variable) (logistic regression) 0. (8) (modeling) (meta-analyses) ( ) (odds ratio OR) (7,8,9) ( ) (relative risk reduction ) (7,8,9) (effect) 0% (baseline risk) (magnitude of risk without therapy) (decision analysis) (control event rate CER) % RCTs CER baseline risk CER %

5 24 50% ( ) (absolute risk reduction, ) (7,8,9) 50% (risk difference) (attributable risk reduction) OR. 2 x 2 Controlled Event Rate (CER) Experimental Event Rate (EER) c c+d a a+d Relative Risk (RR) a/(a+b) c/(c+d) a c Outcome ( )* b d OR Relative Risk Reduction () -RR c/(c+d)-a/(a+b) c/(c+d) c a Absolute Risk Reduction () - c+d a+d Number Needed to Treat () a/b ad Odds Ratio (OR) c/d cb risk difference (effect size) (0) *. (event rate) () /= (CER) (EER) %.2 (.0,.6)* /0.2=8 (6, 0)* %.06 (.05,.08)* /0.06=7 (2.5, 20)* * 95% confidence intervals 95% Confidence Interval (CI) on an = / (limits on the CI of its ) = +/.96 * (Square Root of ((((CER*(-CER))/No. of Control Pts.) + (((EER*(-EER))/No. of Experimental Pts.)))

6 25 ( ) (number needed to treat ) (7,8,9) () (NNH) 00mg NNH 00mg (9,2) Laupacis (9) number needed to be treated number needed to treat for benefit for harm for benefit for harm for benefit number needed to harm (NNH) NNH (9) A B 95% (decision analysis) () NNH (absolute risk increase ARI) (4) (ARI) ARI (relative risk increase RRI) ARI RRI () (2,3) benefit harms (co-morbidity) Cook Sackett (adjusted ) (2) f adjusted f (9,2,5) (RCTs) (fol-

7 26 low-up period) (time-dependent) () (a constant over time) Laupacis T S (T) T S (S) (9) 0% (S) Chatellier (6) CER (5,7). Nuovo J, Melnikow J, Chang D. Reporting number needed to treat and absolute risk reduction in randomized controlled trials. JAMA 2002;287: Anonymous. Evidence-based medicine: a new approach to teaching the practice of medicine. Evidence-Based Medicine Working Group. JAMA 992;268: Begg C, Cho M, Eastwood S, et al. Improving the quality of reporting of randomized controlled trials: The CON- SORT statement. JAMA 996;276: Altman DG, Schulz KF, Moher D, et al. The revised CON- SORT statement for reporting randomized trials: explanation and elaboration. Ann Intern Med 200;34: Junker CA. Adherence to published standards of reporting: a comparison of placebo-controlled trials published in English or German. JAMA 998;280: Moher D, Jones A, Lepage L. Use of the CONSORT statement and quality of reports of randomized trials a comparative before-and-after evaluation. JAMA 200;285: Straus SE, Richardson WS, Glasziou P, Haynes RB. Chapter 5: Therapy. Evidence-Based Medicine: How to Practice and Teach EBM. 3rd ed. London: Elsevier, 2005: Jaeschke R, Guyatt G, Barratt A, et al. Therapy and understanding the results: measures of association. In: Guyatt G, Rennie D, ed. User s Guides to the Medical Literature. Chicago: AMA Press, 2002: Laupacis A, Sackett DL, Roberts RS. An assessment of clinically useful measures of the consequences of treatment. N Engl J Med 988;38: Newcombe RG. Confidence intervals for the number needed to treat: absolute risk reduction is less likely to be misunderstood. BMJ 999;38: Straus SE, Richardson WS, Glasziou P, et al. Appendix : Confidence intervals. Evidence-Based Medicine: How to Practice and Teach EBM. 3rd Ed. London: Elsevier, 2005: Cook RJ, Sackett DL. The number needed to treat: a clinically useful measure of treatment effect. BMJ 995;30: Walter SD. Number needed to treat (): estimation of a measure of clinical benefit. Stat Med 200;20: Smith GD, Egger M. Who benefits from medical interventions? BMJ 994;308: Tramer MR, Walder B. Number needed to treat (or harm). World J Surg 2005;29: Chatellier G, Zapletal E, Lemaitre D, et al. The number needed to treat: a clinically useful nomogram in its proper context. BMJ 996;32: Wu LA, Kottke TE. Number needed to treat: caveat emptor. J Clin Epidemiol 200;54:-6.